This paper conducts a reconstruction analysis of the intelligent maintenance system for power transmission equipment from the perspectives of digital twins and artificial intelligence. After constructing the intelligent maintenance system, the technical architecture of the system is established using technologies such as multi-source heterogeneous databases. Additionally, the paper explores the integration of artificial intelligence technology into the maintenance management of power systems to achieve precise assessments of the system’s health status and tests the performance of the proposed evaluation method. In terms of system health assessment, the relative error of the proposed method does not exceed 0.012, with an average assessment accuracy of 99.39% and an average assessment time of 150.63 seconds. Among all assessment methods, the proposed method achieves the highest accuracy and fastest speed, demonstrating significant advantages in power equipment health assessment.